Afm Spring Adjustment - Bothersome Thoughts

[Zed Head]

I haven't been able to really define what it is about adjusting the spring of the AFM that bothers me.  But I think I've narrowed to down to one thing - preload.  

 

Preload affects when the vane starts to move due to air flow and should be an exact number, matched to the idle air bypass in the AFM and the throttle blade idle air bypass.  Any adjustment of the AFM spring wheel will affect that three way relationship.  That's probably why the factory glues the wheel down when they've adjusted it.  But preload does not affect spring rate so once the vane starts moving the change in injector time versus vane position goes to the original design goals.

 

If you tighten the spring, the engine gets extra air before the vane moves and enrichment starts.  But it's only a certain amount of low flow air.  So you'll be leaner at the low air flow than at high air flow.

 

If you loosen the spring you'll get extra enrichment at low air flow, but the effect will diminish at higher air flow.  So you'll be richer at low air flow but little effect at high air flow.

 

The spring is a spiral spring with essentially constant spring rate over the range of movement in the application.  So, unlike the potentiometer on the coolant temperature sensor modification, the AFM spring adjustment really only has significant effect at low air flow conditions.  The potentiometer affects the full range of conditions.

 

Springs are weird things that are pretty simple until you try to figure out how to use them.

[siteunseen]

[Chickenman]

 

But preload does not affect spring rate so once the vane starts moving the change in injector time versus vane position goes to the original design goals.

 

That statement is correct, as far as Spring Rate,  but you may have taken it out of context in relation to how the Spring Load curve affects AFM flap operation. Spring Load is a factor of Spring Rate + Spring Pre-load + Spring Compression Height.

 

Spring pre-load doesn't affect Spring Rate...but it does increase total spring resistance or Spring Load at a given point.

 

Take a typical coil spring that is rated at 100 lbs/in and throw it spring a spring tester. At 0"  compression height there is zero resistance ( no load condition ). At 1" of compression you should have 100 lbs of resistance. At 2" compressed height you should have 200 lbs of resistance. At 3" compression you would have 300 lbs resistance...and so on.

 

Now add 1" of pre-load to the spring. Reset the spring height pointer to zero.  At 0" indicated compression height  you would have 100 lbs of resistance. At 1" compression height you would have 200 lbs of resistance. At 2" compression height you would have 300 lbs of resistance.... and so on.

 

So increasing spring pre-load on your AFM spiral spring is going to change spring resistance or Load in a linear curve, throughout the entire sweep range.

Edited March 23, 2015 by Chickenman

[Chickenman]

And if memory serves me correct the AFM only provided mixture measurement and correction up to about 4,500 rpm. After that the ECU calculates IDC entirely based on RPM, TPS signal and CTS.

[FastWoman]

I can't imagine this is quite true, although we've all certainly heard it.  There's a big range of throttle positions between "off idle" and not quite "WOT," per the TPS signal.  Are we to assume an engine spinning 4500 or above will get the same amount of fuel, regardless, metered only in proportion to RPM?  I doubt it -- although again, we've all heard it.

 

What bothers me about monkeying with the clock spring is that this not only impacts the preload, but also impacts the airflow at which the AFM pegs out (which it does).  If we loosen the spring to deliver more fuel, the vane pegs out with less airflow, and then there is no room to deliver more, except as RPM increases.  This results in an engine leaning out at the very time it is worst for it to lean out -- under heavy load, revving hard.

[Chickenman]

It really doe's work that way FW.  There is a  correlation between the TPS switch and engine RPM. The TPS has WOT contacts and goes to a different Map ( Full Power rich ) when it sees the WOT switch close. Fuel injector duty cycle is relatively easy to calculate at WOT and varying RPM's. It's varying loads at part throttle and mid-rpm's that get tricky.

 

Edit: Don't forget, at 4,500+ RPM you are nearly always in a WOT, hard acceleration condition. You don't cruise very often at 4,500 RPM or vary the throttle much. Anything above 4,500 RPM is a Full Power Mode and  wants around 12.5 to 1 ( +/- .5 AFR ) and that is fairly constant. Everything works around BSFC and once you set that, you really do just increase the Fuel amount based on RPM. This on Normally Aspirated engines of course. Forced Induction needs Boost compensation as well.

 

Edit 2: The AFM meter maxing out early will have no effect on supplying more fuel at high RPM's and WOT.  The ECU ( on our cars ) will be in " Open Loop " under those conditions and will not be even reading the AFM values. Engines at WOT, heavy load and above approx 4,000 RPM all  all need a Full Power Enrichment AFR ratio of about 12 to 12.5 under those conditions. A simple full power fuel map can calculate the IDC based on rising RPM's to maintain that calculated AFR.

 

Vane type AFM's, on all cars, become less sensitive as the flap nears fully open. They start losing hysteresis. For that reason, manufacturers go to an RPM based Fuel/Timing Map once the flap reaches 100%,  TPS indicates WOT,  and RPM rises above a certain point.

 

You can have VERY accurate fuel mapping done with just a TPS switch, Coolant sensor switch, and RPM sensing. Throw in a MAP ( Manifold Absolute Pressure ) switch into the equation and you have a Speed Density System which works very well on cars with big cams. GM used it successfully on it's 86 to 1989  TPI system. They only switched to a MAF measuring system around 1990.

 

Don Devendorf developed 900 HP on the Electromotive 280ZX Turbos back in the early 1980's with a Speed Density System that used no Airflow meters. And that was on a Turbo with varying boost levels. A Normally Aspirated motor is much easier to calculate IDC and fuel curves than a Turbo motor. .

 

Vane type airflow meters were never accurate enough at high airflow rates to safely calculate IDC by themselves. MAF meters with hot film sensors changed that and are a lot more accurate, but even then, at WOT and high RPM some fairly modern  1990's ECU's go into Open Loop with the MAF sensor and NB O2 sensor cut out of the loop and the ECU relies on a set " Power Map " for fuel and timing.

 

MAF sensors ( Hot Film style ) and WB O2 sensors  allow the ECU to stay in Closed Loop all of the time. Even then, some Manufacturers will still go to " Open Loop " at WOT. It's a primarily an Engine Safety Strategy.

Edited March 24, 2015 by Chickenman

[Chickenman]

Even fairly modern systems such as the Bosch ME5.2 ( Narrow Band O2 sensor )   go into " Open Loop " at WOT above a certain RPM. The MAF sensor has no effect in " Open Loop " .  It's a very common strategy in ECU mapping. Hitachi and Nippon-Denso do the same thing with Japanese ECU's.

 

Narrow Band O2 sensors have a very limited range. They are only accurate in a narrow range close to Stoich ( 14.7 to 1 AFR ). The NB O2 sensor cannot measure mixtures accurately enough under hard throttle, high RPM's and Boost. The richer " Power " mixtures are outside of the sensors measurement range. So the ECU's just goes " Open Loop " and relies on a set fuel Mapping to calculate IDC. No input from the MAF sensor at all. 

 

My 1998 AEB engined Audi A4 1.8T  does not even have a MAP sensor. ( Stock design ). It only has a MAF and NB O2 sensor. At WOT the ECU goes Open loop ( Speed Density )  by design and doesn't even use the MAF input. Yet it runs very happily to 7,200 rpm with about 14 lbs boost ( 18 lbs at 6,000 RPM ). And that's with only a Stage 1 chip.   Drag Racers on some of the forums I frequent have made over 900 HP out of VW/Audi engines, using stand alone ECU management. Speed Density systems with no MAF sensors or AFM's at all. 

 

Modern F1 engines from the early 1980's through current 2014 have never used any type of Airflow Meter. All fuel maps are made based on MAP sensors, RPM's and TPS position. That's all the input they need. Race engines are nearly always at WOT. And WOT is very easy to calculate fuel flows. All using Speed Density systems.

Edited March 23, 2015 by Chickenman

[240260280z]

The analog nature of the 280z ECU would require a frequency reference or frequency to voltage reference and a comparator tuned to flip at 4500 rpm for the input from the AFM flap. It would be interesting to see if this exists in the circuitry.

[Chickenman]

The analog nature of the 280z ECU would require a frequency reference or frequency to voltage reference and a comparator tuned to flip at 4500 rpm for the input from the AFM flap. It would be interesting to see if this exists in the circuitry.

I think the ECU Bible or one of the articles by " Brapp " ( ??? ) mentions the frequency counter and voltage reference. Should be dead easy to incorporate into a circuit design. 

 

I believe that " Brapp " and the Bible also mention the 4,500 to 4,800 rpm limit for the Vane style AFM.  I'll look in a couple of days for the link when I have more time....

 

 

I've got some books on modding Fuel Injection systems and one of the Project cars was a 1991  Toyota MR2 Turbo ( Factory stock ). Same style of Analog Vane AFM that we have. Seriously hard to modify when Boost is increased, but there are specialists who do it.

 

One of the main issues ( On the Turbocharged)  MR2  was that the ECU was Analog and it switched to Speed Density once RPM reached a certain point. AFM again was maxed out at around 4,500 RPM or earlier with more boost than stock. Analog is just so difficult to work with. You can't re-program Cells like a Digital ECU. It's all about tweaking resistors, Pots and other such nonsense. 

 

Much easier option is to go with a modern Digital Stand-Alone with Wide Band and fully programmable, but it was interesting to see just how much they could re-engineer the OEM Analog system. They did some pretty serious Mods to that poor Mother Board !!

 

Modifications in article went all the way from Stock Analog ECU with Vane style AFM, to a VPC ( Vane Pressure Convertor ) that eliminates the restrictive and troublesome Vane style AFM to a MAP sensor style Piggy back with a built in Microprocessor that can be programmed to change the fuel curve. They even went as far as converting the original Analog motherboard with a Plug-in Daughter board to make the factory ECU digitally programmable. Eventually, after much fiddling and farting about and raising of Power levels...they ditched the whole mess and went to a Motec 48 stand-alone EMS... which is what they should have done in the first place, IMHO. It was an interesting read though.... 16 Stages that went from stock 165.2 RWHP to eventual 471 RWHP with a 2.1 stroker and Motec 48 EMS. And documented all the trials and tribulations.... all 21 pages!!

 

Edit: Book is " How to Tune and Modify Engine Management Systems " by Jeff Hartman. ISBN 0-7603-1582-5 . A great technical book on Modding old and new EFI systems. It's a good read...

Edited March 24, 2015 by Chickenman

[mjr45]

Chickenman, you and Blue are way over my head in this post, I know spark, air and fuel much past that and I start getting lost, but this an interesting read and informative for dummy's like me.

[Mark Maras]

 Ditto, Thanks.

[240260280z]

I have heard about this "4500" limit but don't think a 4500 rpm limit exists until someone shows me the circuit and its purpose. I can rev to 4500 out-of-gear with very little air by hand reving or I can cruise down the highway in 3rd and hold at 4500 or I rev to 4500 WOT in 5th tearing down a straight at Bonneville. The CFM of air entering the engine is different for each 4500 and the AFM Vane will be in different locations.

It is a bit of circuit de-constructing that is required to figure it out for ourselves. Even the Nissan EFI fuel table (below) does not show the AFM contribution! So there are a lot of unknowns.

Edited March 24, 2015 by Blue